Azimuth sound localization in many animals relies on the processing of differences in time-of-arrival of the low-frequency sounds at both ears: the interaural time differences (ITD). It was observed in some species that this cue depends on the spectrum of the signal emitted by the source. Yet, this variation is often discarded, as humans and animals are assumed to be insensitive to it. The purpose of this thesis is to assess this dependency using acoustical techniques, and explore the consequences of this additional complexity on the neurophysiology and psychophysics of sound localization. In the vicinity of rigid spheres, a sound field is diffracted, leading to frequency-dependent wave propagation regimes. Therefore, when the head is modeled as a rigid sphere, the ITD for a given position is a frequency-dependent quantity. I show that this is indeed reflected on human ITDs by studying acoustical recordings for a large number of human and animal subjects. Furthermore, I explain the effect of this variation at two scales. Locally in frequency the ITD introduces different envelope and fine structure delays in the signals reaching the ears. Second the ITD for low-frequency sounds is generally bigger than for high frequency sounds coming from the same position. In a second part, I introduce and discuss the current views on the binaural ITD-sensitive system in mammals. I expose that the heterogenous responses of such cells are well predicted when it is assumed that they are tuned to frequency-dependent ITDs. Furthermore, I discuss how those cells can be made to be tuned to a particular position in space irregardless of the frequency content of the stimulus. Overall, I argue that current data in mammals is consistent with the hypothesis that cells are tuned to a single position in space. Finally, I explore the impact of the frequency-dependence of ITD on human behavior, using psychoacoustical techniques. Subjects are asked to match the lateral position of sounds presented with different frequency content. Those results suggest that humans perceive sounds with different frequency contents at the same position provided that they have different ITDs, as predicted from acoustical data. The extent to which this occurs is well predicted by a spherical model of the head. Combining approaches from different fields, I show that the binaural system is remarkably adapted to the cues available in its environment. This processing strategy used by animals can be of great inspiration to the design of robotic systems.

The Short-Time Target Cancellation (STTC) algorithm, developed as part of this dissertation research, is a “Cocktail Party Problem” processor that can boost speech intelligibility for a target talker from a specified “look” direction, while suppressing the intelligibility of competing talkers. The algorithm holds promise for both automatic speech recognition and assistive listening device applications. The STTC algorithm operates on a frame-by-frame basis, leverages the computational efficiency of the Fast Fourier Transform (FFT), and is designed to run in real time. Notably, performance in objective measures of speech intelligibility and sound source segregation is comparable to that of the Ideal Binary Mask (IBM) and Ideal Ratio Mask (IRM). Because the STTC algorithm computes a time-frequency mask that can be applied independently to both the left and right signals, binaural cues for spatial hearing, including Interaural Time Differences (ITDs), Interaural Level Differences (ILDs) and spectral cues, can be preserved in potential hearing aid applications. A minimalist design for a proposed STTC Assistive Listening Device (ALD), consisting of six microphones embedded in the frame of a pair of eyeglasses, is presented and evaluated using virtual room acoustics and both objective and behavioral measures. The results suggest that the proposed STTC ALD can provide a significant speech intelligibility benefit in complex auditory scenes comprised of multiple spatially separated talkers. / 2020-10-22T00:00:00Z

Binaural beamforming technology, which is based on the auditory perception of both ears, uses a wireless data connection to exchange data between the right-side and the left-side hearing aids. Over the years, several multichannel speech enhancement algorithms have been used in the hearing aid industry. For example, beamforming algorithms work by keeping a target signal undistorted while attenuating the noise fields (such as diffuse noise or white noise) and the interferers from different directions. Fixed and adaptive algorithms of this nature have been under active investigation by the hearing aid industry. Although binaural beamforming hearing aids designs have shown better performance than single-channel based hearing aids or bilateral hearing aids, the performance of binaural beamforming still suffers from errors in the direction of arrival estimates, i.e., errors which occur when the right set of steering vectors is used in a beamformer design but the target signal source is not located at the direction considered in the design. Therefore, this thesis is devoted to find and propose structures showing more robustness to errors in the direction of arrival estimates. The focus is mainly on the Generalized Sidelobe Canceller (GSC) structure and several binaural beamforming algorithms and configurations are proposed in this thesis as alternatives for the fixed beamformer and blocking matrix units of the GSC. The proposed algorithms show promise of providing wider notch and/or wider beam possibilities, as well as providing greater noise reduction and superior adaptive null positioning capabilities. The algorithms proposed in this thesis were simulated in MATLAB using recorded signals and data provided by a hearing aid firm, to assess their utility for improving hearing aid performance. The results demonstrated a superiority over algorithms currently in use in industry.

There are no data available in the literature that have specifically evaluated differences in adaptation to unilateral conductive or sensory changes. However, based on clinical experience it may be postulated that changes of outer or middle ear function appear to be tolerated more easily than those of cochlear origin. Very often, patients seen in the clinic are unaware of a slight conductive hearing loss. By contrast, patients are immediately disturbed by a minor decline of cochlear function. One of several complaints of these patients is a change in their spatial orientation or difficulties in understanding speech in a noisy environment . The goal of the study was to determine if binaural performance tested psychoacoustically using a lateralization task is influenced differentially by short-term induced unilateral conductive or sensory changes. Lateralization performance was evaluated in seven normal-hearing subjects during induced auditory periphery asymmetry resulting from: l. exposure to noise presented for 5 minutes at 115 dBA SPL or 2. bilateral occlusion with earplugs of unequal attenuation for 48 hrs. An adaptive procedure was used to detenmine hearing thresholds of a 4-kHz narrow-band noise (NBN). In a lateralization task subjects indicated the positions of intracranial images created by the same NBN pr esented binaurally at SO dB SL with interau ral level differences ( I LDs) varying within plus/minus 12 dB. The tests were performed over a one-hour period post-exposure, immediately prior to and following plugging the ears, and at 24 and 48 hrs post-plugging. Immediately after the exposure or after plugging, there was a shift of lateral ization towards an unexposed side or the side blocked by the plug with a smaller attenuation, respectively. After a few minutes post-exposure, signals with I LD=0 were lateralized at midl ine. Within 30 minutes post -plugging, those signals were gradually lateralized closer to midline but remained off center for the rest of the plugging period. Thus, subjects showed fast adaptation to induced unilateral sensorineural changes and incomplete adaptation to induced asymmetrical conductive changes. Those rather unexpected r esults can be explained using a qualitative model assuming that: 1. a conductive impainment reflects a loss of sensitivity and 2. a cochlear impairment reflects both a loss of sensitivity and of the compressive nonlinearity on the basilar membrane. Recently, there has been an increase in the number of psychoacoustical studies on hearing-impaired listeners with a majority of them directed toward revealing deficits in monaural processing. However, in most acoustic environments encountered in everyday life, there are multiple sounds originating from different sources, and hearing-impaired people often display less binaural advantage than do normally hearing persons. The results of the cu rrent study support the view of the lack of a simple relationship between monaural and binaural processing, which is often r eported in studies on hearing-impaired people. This is an important issue in the process of fitting hearing aids binaurally.

Many cells in the inferior colliculus (IC) are excited by contralateral and inhibited by ipsilateral stimulation and are thought to be important for sound localization. These excitatory-inhibitory (EI) cells comprise a diverse group, even though they exhibit a common binaural response property. Previous extracellular studies proposed specific excitatory and/or inhibitory events that should be evoked by each ear and thereby generate each of the EI discharge properties. The proposals were inferences based on the well established response features of neurons in lower nuclei, the projections of those nuclei, their excitatory or inhibitory neurochemistry, and the changes in response features that occurred when inhibition was blocked. Here we recorded the inputs, the postsynaptic potentials, discharges evoked by monaural and binaural signals in EI cells with in vivo whole cell recordings from the inferior colliculus (IC) of awake bats. We also computed the excitatory and inhibitory synaptic conductances from the recorded sound evoked responses. First, we showed that a minority of EI cells either inherited their binaural property from a lower binaural nucleus or the EI property was created in the IC via inhibitory projections from the ipsilateral ear, features consistent with those observed in extracellular studies. Second, we showed that in a majority of EI cells ipsilateral signals evoked subthreshold EPSPs that behaved paradoxically in that EPSP amplitudes increased with intensity, even though binaural signals with the same ipsilateral intensities generated progressively greater spike suppressions. These ipsilateral EPSPs were unexpected since they could not have been detected with extracellular recordings. These additional responses suggested that the circuitry underlying EI cells was more complex than previously suggested. We also proposed the functional significance of ipsilaterally evoked EPSPs in responding to moving sound sources or multiple sounds. Third, by computing synaptic conductances, we showed the circuitry of the EI cells was even more complicated than those suggested by PSPs, and we also evaluated how the binaural property was produced by the contralateral and ipsilateral synaptic events. / text

Στα πλαίσια της διπλωματικής αυτής, έγινε ανάλυση και υλοποίησης της κωδικοποίησης αμφοιωτικών παραμέτρων (BCC-Binaural Cue Coding). Οι αλγόριθμοι της κωδικοποίησης BCC και γενικά όλες οι συναρτήσεις που χρησιμοποιήθηκαν έγιναν στο περιβάλλον της MATLAB. Για τη διαδικασία και για τον έλεγχο των αλγορίθμων έγινε χρήση στερεοφωνικών σημάτων τα οποία κωδικοποιήθηκαν σε μονοφωνικά. Στόχος της κωδικοποίησης ήταν να επιτευχθεί γρήγορη μίξη και εξαγωγή των παραμέτρων του στερεοφωνικού σήματος και ταυτόχρονα όσο το δυνατό πιο πειστική επαναδημιουργία του σήματος. Για αυτό τον λόγο στην προσπάθεια δημιουργίας προγράμματος με γρήγορη απόκριση, η κωδικοποίηση έγινε, τόσο στο πεδίο του χρόνου όσο και στο πεδίο της συχνότητας. / --

Cette thèse concerne la définition d'algorithmes pour la localisation de sources sonores (statiques ou mobiles) depuis un capteur binaural mobile en robotique. L'objectif est de développer des stratégies actives, qui combinent les signaux gauche-droite perçus et les ordres moteurs du capteur de façon à compenser les limitations usuelles dans le cas d'un monde statique : levée d'ambiguïtés (par exemple, avant-arrière), récupération de l'observabilité de certaines variables inobservables, etc. L'étude est focalisée sur des stratégies en deux étapes : (1) extraction d'information spatiale et détection d'activité relative à la/les sources par une analyse court-terme des flux audio ; (2) assimilation temporelle de ces données et fusion avec les ordres moteurs du capteur dans un schéma de filtrage stochastique.

La grande majorité des systèmes perceptifs proposés en robotique héritent d'une conception passive de la perception dans laquelle la génération d'une commande motrice est l'étape ultime d'une succession de traitements purement passifs. Dans le cadre de la localisation de sources sonores, qui est une tâche fondamentale du système auditif, cette approche passive offre de bons résultats lorsque les conditions environnementales sont bien connues et facilement modélisables. Cependant des difficultés apparaissent lorsque l'environnement se complexifie, a fortiori s'il est inconnu ou changeant. Ces difficultés constituent un enjeu important dans le domaine de l'audition artificielle. Cette thèse considère une approche radicalement différente de l'approche passive, inspirée de la psychologie de la perception et de la théorie des contingences sensorimotrices. Cette approche place l'action au coeur du processus de perception, qui est alors vu comme une interaction qu'un agent biologique ou robotique entretient avec son environnement. Alors que l'approche passive nécessite des connaissances sur l'environnement, implicement intégrées dans les traitements par le roboticien, l'approche sensorimotrice suggère au contraire que ces connaissances sont acquises par l'agent de manière autonome, à travers son expérience sensorimotrice. Ainsi cette thèse applique la théorie des contingences sensorimotrices à la localisation de sources sonores pour la robotique autonome. Sur la base d'un modèle bioinspiré du système auditif adapté au contexte robotique, cette thèse propose une redéfinition du problème de la localisation en termes sensorimoteurs. Un modèle de localisation sensorimotrice est alors proposé. Celui-ci se base sur des capacités de perception active bas-niveau pour construire une représentation de l'espace auditif qui est ensuite utilisée pour une localisation passive. En exploitant les capacités d'action du robot, ce modèle permet de s'affranchir des dépendances à l'environnement qui mettent en difficulté l'approche passive, en proposant ainsi un degré d'autonomie supérieur à celui des modèles actuels

The acoustic scattering properties of the human head and torso match well with those of simple geometric shapes. Consequently, analytical scattering models can be utilised to account for the sound localisation cues introduced by these features. The traditional use of such models assumes that the head surface is completely rigid in nature. This thesis is concerned with modelling and understanding the effect of terminal scalp hair (i.e., a non-rigid head surface) on the auditory localisation cues. The head is modelled as a sphere, and the acoustical characteristics of hair are modelled using a locally-reactive equivalent impedance parameter. This allows the scattering boundary to be defined on the inner rigid surface of the head. The boundary assumptions are validated experimentally, through impedance measurement at oblique incidence and analysis of the near-field scattering pattern of a uniformly covered sphere. The impedance properties of human hair are also discussed, including trends with variations in sample thickness, bulk density, and fibre diameter. A general solution for the scattering of sound by a sphere with an arbitrarily distributed, locally reactive surface impedance is then presented. From this, an analytical solution is derived for a surface boundary that is evenly divided into two uniformly distributed hemispheres. For this boundary condition, cross-coupling is shown to exist between incoming and scattered wave modes of equi-order when the degrees are non-equal and opposite in parity. The overall effect of impedance on the resultant scattering characteristics is discussed in detail, both for uniform and for hemispherically divided surface boundaries. Finally, the analytical formulation and the impedance characteristics of hair are collectively utilised to investigate the effect of hair on human auditory localisation cues. The hair is shown to produce asymmetric perturbations to both the monaural and binaural cues. These asymmetries may help to resolve localisation confusions between sound stimuli positioned in the front and rear hemi-fields. The cue changes in the azimuth plane are characterised by two predominant features and remain consistent regardless of the decomposition baseline (i.e., the inclusion of a pinna offset, neck, etc). Experimental comparisons using a synthetic hair material show a good agreement with simulated results.

At a cocktail party, one can choose to scan the room for conversations of interest, attend to a specific conversation partner, switch between conversation partners, or not attend to anything at all. The ability of the normal-functioning auditory system to flexibly listen in complex acoustic scenes plays a central role in solving the cocktail party problem (CPP). In contrast, certain demographics (e.g., individuals with hearing impairment or older adults) are unable to solve the CPP, leading to psychological ailments and reduced quality of life. Since the normal auditory system still outperforms machines in solving the CPP, an effective solution may be found by mimicking the normal-functioning auditory system. Spatial hearing likely plays an important role in CPP-processing in the auditory system. This thesis details the development of a biologically based approach to the CPP by modeling specific neural mechanisms underlying spatial tuning in the auditory cortex. First, we modeled bottom-up, stimulus-driven mechanisms using a multi-layer network model of the auditory system. To convert spike trains from the model output into audible waveforms, we designed a novel reconstruction method based on the estimation of time-frequency masks. We showed that our reconstruction method produced sounds with significantly higher intelligibility and quality than previous reconstruction methods. We also evaluated the algorithm's performance using a psychoacoustic study, and found that it provided the same amount of benefit to normal-hearing listeners as a current state-of-the-art acoustic beamforming algorithm. Finally, we modeled top-down, attention driven mechanisms that allowed the network to flexibly operate in different regimes, e.g., monitor the acoustic scene, attend to a specific target, and switch between attended targets. The model explains previous experimental observations, and proposes candidate neural mechanisms underlying flexible listening in cocktail-party scenarios. The strategies proposed here would benefit hearing-assistive devices for CPP processing (e.g., hearing aids), where users would benefit from switching between various modes of listening in different social situations. / 2022-05-19T00:00:00Z